No kidding.


Jonathan Gray <jsg@jsg.id.au> wrote:

> On Fri, Aug 30, 2024 at 11:54:31AM +0200, Alexander Bluhm wrote:
> > On Thu, Aug 29, 2024 at 11:15:55AM +0200, Hans-J?rg H?xer wrote:
> > > Hi,
> > > 
> > > On Thu, Aug 29, 2024 at 10:14:12AM +1000, Jonathan Gray wrote:
> > > > On Thu, Aug 29, 2024 at 09:28:50AM +1000, Jonathan Gray wrote:
> > > > > >  
> > > > > > +#if NCCP > 0
> > > > > > +#if NVMM > 0
> > > > > 
> > > > > can't this be only #if NCCP > 0?
> > > 
> > > yes, there's actually no need for depending on NVMM
> > 
> > RAMDISK_CD does not build.  It has ccp, but #include <machine/conf.h>
> > is within #if NVMM > 0.
> > 
> > /crypt/home/bluhm/openbsd/cvs/src/sys/kern/kern_pledge.c:1357:44: error: use of undeclared identifier 'pspopen'; did you mean 'pppopen'?
> >                     (cdevsw[major(vp->v_rdev)].d_open == pspopen)) {
> >                                                          ^~~~~~~
> >                                                          pppopen
> > 
> > If there is no VMM, the psp call make not much sense.  So I put
> > back #if defined(__amd64__) && NCCP > 0 && NVMM > 0
> 
> I'd prefer it to be detangled
> 
> Index: kern_pledge.c
> ===================================================================
> RCS file: /cvs/src/sys/kern/kern_pledge.c,v
> diff -u -p -U4 -r1.317 kern_pledge.c
> --- kern_pledge.c	1 Sep 2024 17:13:46 -0000	1.317
> +++ kern_pledge.c	2 Sep 2024 00:10:38 -0000
> @@ -72,12 +72,10 @@
>  #include "pty.h"
>  
>  #if defined(__amd64__)
>  #include "vmm.h"
> -#if NVMM > 0
> -#include <machine/conf.h>
> -#endif
>  #include "ccp.h"
> +#include <machine/conf.h>
>  #endif
>  
>  #include "drm.h"
>  
> @@ -1349,9 +1347,9 @@ pledge_ioctl(struct proc *p, long com, s
>  		}
>  	}
>  #endif
>  
> -#if defined(__amd64__) && NCCP > 0 && NVMM > 0
> +#if NCCP > 0
>  	if ((pledge & PLEDGE_VMM)) {
>  		if ((fp->f_type == DTYPE_VNODE) &&
>  		    (vp->v_type == VCHR) &&
>  		    (cdevsw[major(vp->v_rdev)].d_open == pspopen)) {
>